As spoken → as typed
So I was thinking we should, uh, ship on Monday, no wait, Tuesday, because the the tests are not done.
removed by Saylo, on your computer, before anything is typed
[01] Local voice dictation
Talk. Saylo types it for you.
Voice dictation for Mac and Windows. Hold a key, speak, let go. The finished sentence appears in whatever app you’re using — with the “ums” removed and your corrections applied. It all runs on your own computer.
Free · macOS 12+ and Windows 10/11 · No account
[02] In Slack
P
Priya
Is standup still at 10 tomorrow?
A
You
Can we move the standup to 11 tomorrow?
Spoken: “Can we, um, move the standup to 10, no wait, 11 tomorrow?”
Works offline
No account, ever
Ready the second it opens
Free, no word limits
MailSlackNotionGoogle DocsVS CodeTerminalMessagesLinearPagesChromeCursorOutlook
[04] How it works
Three moves. Nothing to open.
-
01
Right ⌥
Hold the key
Press and hold Right Option (Right Alt on Windows) anywhere. A small pill appears so you know Saylo is listening.
-
02
Say it
Talk naturally. Your words appear in the pill while you speak, so you know it heard you.
-
03
Ship the notes by Friday
Let go
Release the key and the clean sentence is typed where your cursor is.
Prefer hands-free? Tap the key once to start and again to finish. Press Esc to cancel.
[05] What it does
It types what you meant, not what you said.
Nobody speaks in clean prose. Saylo tidies the transcript on your computer, in a fraction of a second, before a single character is typed.
[06]
Fillers removed
“Um”, “uh”, “you know” and stutters like “the the” never reach the page.
[07]
Corrections applied
Say “Friday, actually Thursday” or “scratch that”, and only what you meant is typed.
[08]
Works offline
Speech recognition runs on your computer. Turn Wi-Fi off and keep dictating.
[09]
Voice commands
Punctuation is added for you. Or say it: “new line”, “new paragraph”, “comma”, “question mark”.
[10] New in 0.1.5 · Paragraphs
Long dictations come out in paragraphs
Breaks are added at sentence ends in long dictations — never in chat apps, terminals or code editors.
[11]
See it as you speak
Words appear in the pill while you’re still talking.
[12]
Your vocabulary
Add names and jargon once. Saylo hands them to the speech model and fixes near-misses — “Superbase” becomes “Supabase”.
[13]
Dozens of languages
English, Hindi, Spanish, French, German, Japanese, Chinese and more. Pick one, or let Saylo detect it.
Right ⌥
One key, any app
Hold to talk, tap for hands-free, Esc to cancel. Or use Right ⌘, Right Control, F5 or F6.
[14]
Searchable history
Find and copy earlier dictations, with your words per minute. Kept in a file on your computer — or switched off.
[15]
Nothing to set up
A speech model is built in, so the first press works. Saylo then fetches a better one for your machine in the background.
[16]
Light when idle
Models unload after a few quiet minutes and load again on your next dictation. Silence is skipped before transcription.
[17] Context
It knows which app you’re typing into.
Dictation that formats a shell prompt like an email is worse than none. Saylo checks what’s in front before it types.
- Terminals and code editors get your words exactly — no capitals or paragraph breaks added.
- Chat apps get one clean message, without paragraph breaks.
- Continuing a sentence? Saylo matches the text before your cursor, so it starts lowercase and adds the space.
- Names on screen in the window title are spelled right (Mac).
One switch turns app-awareness off.
Terminal
can you add a loading spinner to the settings pageSlackOn my way. Five minutes.
MailHi Sam, thanks for the quick turnaround. I have signed the contract and sent it back.
[18] Inside the app
Settings you’ll open twice, then forget.
Saylo with sample dictations.
[19] Privacy
Your voice never leaves your computer.
Saylo makes only three kinds of network request, and none carries anything about you: downloading a speech model, checking for a new version, and downloading it. Updates can be switched off. Turn Wi-Fi off and keep dictating.
- Audio is processed in memory and discarded — never written to disk, never uploaded
- No account, no in-app analytics, no crash reporting
- The speech engine listens on
127.0.0.1only — nothing is exposed - Sandboxed interface, strict content security policy
Cloud dictationSaylo
Where audio is processedVendor serversYour device
Account requiredYesNo
Works offlineNoYes
Audio retentionPer vendor policyNever stored
Word limitsCaps on free plansNone
PriceTypically $12–15 a monthFree
[20] Models
Chosen for your machine, or chosen by you.
On first launch Saylo checks your processor and memory, picks the speech model that fits and downloads it in the background while you dictate with the built-in one.
| Speech model | Size | Best for | Quality |
|---|---|---|---|
| Base Built in | 60 MB | Ships inside the app — works before you download anything | |
| Small | 190 MB | Chosen on Windows and Intel Macs — runs on the CPU | |
| Large v3 Turbo Recommended | 574 MB | Apple Silicon — near-realtime on the GPU | |
| + Neural Engine Optional | +1.1 GB | Apple Silicon with Large v3 Turbo — text about twice as fast after you stop talking, same accuracy |
[21] Download
Get Saylo. It works the moment it opens.
Free, no account. A speech model is built into the installer, so you can dictate straight away.
Your system
macOS
Universal — Apple Silicon and Intel. macOS 12 or later. Signed and notarized by Apple.
Download for MacFirst launch on macOS
- Open the DMG and drag Saylo into Applications.
- Open Saylo. It’s notarized, so there’s no security warning to click past.
- Allow Microphone and Accessibility when asked. Accessibility lets Saylo hear the hotkey and type into other apps.
- Start dictating. Saylo picks a better speech model for your Mac and downloads it in the background.
Saylo updates itself: new versions download in the background and install when you’re not dictating.
Your system
Windows
64-bit installer for Windows 10 and 11. Hold Right Alt to dictate.
Download for WindowsFirst launch on Windows
- Run the installer. If SmartScreen shows “Windows protected your PC”, click More info → Run anyway.
- Saylo appears in the system tray, ready to use.
- If Saylo can’t hear you, check that Settings → Privacy → Microphone allows desktop apps.
The Windows installer isn’t code-signed yet, which is why SmartScreen asks.
Looking for the changelog? Release notes · Need help? Contact support
On Linux? A build exists but hasn’t been tested on real hardware yet — get in touch if you’d like to try it.
[22] FAQ
Questions people ask before installing.
Is it really offline?
Yes. Turn Wi-Fi off and dictate — cleanup included. Saylo only goes online to download speech models, to check for a new version and to download it, and you can switch updates off. None of these sends anything about you.
Do I need to download anything first?
No. A speech model is built into the installer, so the first press of the hotkey works. Saylo then picks a better model for your computer and switches over when it finishes downloading.
Which apps does it work in?
Any app you can type in: Mail, Slack, browsers, documents, code editors and terminals.
Will it mangle my terminal or code?
In terminals and code editors Saylo types your words exactly, with no capitals or paragraph breaks added. It’s best for prose — messages, prompts and comments. Spoken commands can be misheard (“git” often comes out as “get”), so type exact syntax yourself.
How accurate is it?
With Large v3 Turbo, very good for clear speech in supported languages. Choosing your language in Settings is more accurate and faster than auto-detect, and adding names and jargon to your vocabulary gets them spelled right.
Why does it need Accessibility permission?
To notice the hotkey while another app is in front, and to type the result into that app. Saylo only listens for the key you chose.
Is the Mac app signed and notarized?
Yes. Saylo is signed with an Apple Developer ID and notarized by Apple, so it opens without an “unidentified developer” warning. The Windows installer isn’t code-signed yet, which is why SmartScreen asks.
What hardware do I need?
A Mac running macOS 12 or later, or a 64-bit Windows 10/11 PC. Apple Silicon Macs use the GPU and can use the Neural Engine. Intel Macs and Windows use all CPU cores with a smaller model.
Can I change the hotkey?
Yes: Right Option/Alt, Right Command/Windows, Right Control, F5 or F6. Hold to talk, tap for hands-free, Esc to cancel.
Is my dictation history stored anywhere?
Only in a file on your computer, so you can search it and copy earlier dictations. You can clear it or switch history off. Nothing is synced or uploaded.
Is it free because it’ll start charging later?
There’s no account, no licence check and no server to bill you from, so nothing can be switched on later for the version you’ve installed. Future versions may be paid; the build you download keeps working offline.
Your voice, your computer.
Free, no account, and dictating within a minute of the download finishing.
Download Saylo